We propose to recruit families of the long term survivors in the Cardiovascular Health Study (CHS) for the Multicenter Study on Exceptional Survival in Families (ESF). The CHS cohort (n=5888) was originally recruited from a random sample of Medicare enrollees in 4 US communities in 1989-90. They have been extremely well characterized regarding health and function, providing an opportunity to examine many specific as well as composite phenotypes over time. All have stored DNA or cryopreserved white blood cells. We have determined that about 400 individuals aged 90 and over will be available in CHS, with about 200 remaining free of disability and 100 with fairly large families (average 12) including sibs and their children. We would propose to characterize these family members at several levels of detail, to include the minimal data set suggested by the working group on exceptional survival. Since the disability measures proposed for the minimal data set may well not yet be expressed in the children, we also propose measures that would be [unreadable] informative across a spectrum of age as markers of aging and subclinical disease processes. We have a long history of productive collaboration on large multicenter studies and can contribute to the design of common protocols and analytic approaches. Our goal is to contribute as a study center in the design and implementation of this study. We will recruit and examine older adults who have reached age 90, their siblings and the children of the index generation (index cases and siblings). We will determine the extent of clustering of several potential important exceptional survival phenotype. We can contribute data on environmental and behavioral determinants and intermediate phenotypes that can be assessed in younger old adults and will measure core phenotypic characteristics and important environmental exposures in these families. We will participate in the evaluation of aggregation of selected traits and combination of traits to inform future studies. We would propose first determining whether "exceptional survival" is more common in families than expected and then would use combinatorial techniques (principal component analysis and pedigree discriminant analysis) to produce quantitative phenotypes for genetic analysis. [unreadable] [unreadable]